This study harnesses advanced Bayesian optimisation techniques for the intricate design of structures, poised for application as compliant flooring unit cells, as shown to reduce fall injuries in older people[1]. Inspired by the foundational research detailed in studies[2][3], our exploration delves into a design space encompassing eight key variables, contributing to a broad spectrum of geometric variations. In constitutive modelling of unit cells, we adopt the Neo-Hookean model for its effectiveness in mirroring rubber-like materials’ intrinsic properties. A key element of our approach involves acknowledging and addressing uncertainty, emanating from variations in material properties and geometric imperfections. The study rigorously explores the behaviour of these structures under 2D loading condi- tions, examining various ratios of compression and shear forces. Employing the Bayesian optimisation algorithm, the design iteratively progresses towards an optimal configuration, driven by a comprehensive set of objective functions, which are formulated to achieve an balance between structural stiffness, critical buckling load, and the energy absorption. The effectiveness of the obtained designs is validated by mechanical testing of specimens, manufactured utilizing rubber casting and 3D printing. This study highlights the efficacy of Bayesian optimization in design, particularly under diverse loading conditions and in the presence of uncertainty. It provides valuable in- sights into the method’s ability to improve the resilience and adaptability of structures, representing a notable advancement in the field.